WL#1397 convert XML -> SQL

mysql-test/r/loadxml.result

+drop table if exists t1, t2;
+create table t1 (a int, b varchar(64));
+-- Load a static XML file
+load xml infile '../../std_data/loadxml.dat' into table t1
+rows identified by '<row>';
+select * from t1 order by a;
+a	b
+1	b1
+2	b2
+3	b3
+11	b11
+111	b111
+112	b112 & < > " ' &unknown; -- check entities
+212	b212
+213	b213
+214	b214
+215	b215
+216	&bb b;
+delete from t1;
+-- Load a static XML file with 'IGNORE num ROWS'
+load xml infile '../../std_data/loadxml.dat' into table t1
+rows identified by '<row>' ignore 4 rows;
+select * from t1 order by a;
+a	b
+111	b111
+112	b112 & < > " ' &unknown; -- check entities
+212	b212
+213	b213
+214	b214
+215	b215
+216	&bb b;
+-- Check 'mysqldump --xml' + 'LOAD XML' round trip
+delete from t1;
+load xml infile 'MYSQLTEST_VARDIR/tmp/loadxml-dump.xml' into table t1 rows identified by '<row>';;
+select * from t1 order by a;
+a	b
+111	b111
+112	b112 & < > " ' &unknown; -- check entities
+212	b212
+213	b213
+214	b214
+215	b215
+216	&bb b;
+--Check that default row tag is '<row>
+delete from t1;
+load xml infile 'MYSQLTEST_VARDIR/tmp/loadxml-dump.xml' into table t1;;
+select * from t1 order by a;
+a	b
+111	b111
+112	b112 & < > " ' &unknown; -- check entities
+212	b212
+213	b213
+214	b214
+215	b215
+216	&bb b;
+-- Check that 'xml' is not a keyword
+select 1 as xml;
+xml
+1
+create table t2(fl text);
+LOAD XML LOCAL INFILE "$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml"  INTO TABLE t2 ROWS IDENTIFIED BY '<person>';;
+show processlist;
+Id	User	Host	db	Command	Time	State	Info
+2	root	localhost	test	Query	0	NULL	show processlist
+5	root	localhost	test	Query	3	Reading from net	LOAD XML LOCAL INFILE "$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml"  INTO TABLE t2 ROWS IDENTIFIED BY '<p
+drop table t1;
+drop table t2;
+create table t1 (
+id int(11) not null,
+text text,
+primary key (id)
+) engine=MyISAM default charset=latin1;
+load xml infile '../../std_data/loadxml2.dat' into table t1;
+select * from t1;
+id	text
+1	line1
+line2
+line3
+drop table t1;

mysql-test/std_data/loadxml.dat

+<?xml version="1.0"?>
+<mysqldump xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+<database name="test">
+	<table_structure name="t1">
+		<field Field="a" Type="int(11)" Null="YES" Key="" Extra="" />
+		<field Field="b" Type="varchar(128)" Null="YES" Key="" Extra="" />
+		<options Name="t1" Engine="MyISAM" Version="10" Row_format="Dynamic" Rows="3" Avg_row_length="20" Data_length="60" Max_data_length="281474976710655" Index_length="1024" Data_free="0" Create_time="2007-02-09 09:08:36" Update_time="2007-02-09 09:08:54" Collation="latin1_swedish_ci" Create_options="" Comment="" />
+	</table_structure>
+	<table_data name="t1">
+	<row>
+		<field name="a">1</field>
+		<field name="b">b1</field>
+	</row>
+	<row>
+		<field name="a">2</field>
+		<field name="b">b2</field>
+	</row>
+	<row>
+		<field name="a">3</field>
+		<field name="b">b3</field>
+	</row>
+	<row>
+		<field name="a">11</field>
+		<field name="b">b11</field>
+	</row>
+	
+	<!-- Check field values as tags -->
+	<row>
+		<a>111</a>
+		<b>b111</b>
+	</row>
+
+	<row>
+		<a>112</a>
+		<b>b112 &amp; &lt; &gt; &quot; &apos; &unknown; -- check entities</b>
+	</row>
+
+
+	<!-- Check field values in attributes -->
+	<row a=212 b="b212"></row>
+
+	<!-- Bug#29752 Linefeeds break LOAD XML INFILE -->
+	<!-- Check varios combinations of TAB and NL   -->
+
+	<row
+	a=213 b="b213">
+	</row>
+
+	<row
+	a=214
+	b="b214">
+	</row>
+
+	<row a=215	b="b215"></row>
+
+	<row a=216 b="&bb
+b;"></row>
+
+	<!-- End of bug#29752 -->
+
+	</table_data>
+</database>
+</mysqldump>
+

mysql-test/std_data/loadxml2.dat

+<?xml version="1.0"?>
+<mysqldump xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
+<database name="test_of_xml_import">
+	<table_structure name="t1">
+		<field Field="id" Type="int(11)" Null="NO" Key="PRI" Extra="" />
+		<field Field="text" Type="text" Null="YES" Key="" Extra="" />
+		<key Table="t1" Non_unique="0" Key_name="PRIMARY" Seq_in_index="1" Column_name="id" Collation="A" Cardinality="1" Null="" Index_type="BTREE" Comment="" Index_Comment="" />
+		<options Name="t1" Engine="MyISAM" Version="10" Row_format="Dynamic" Rows="1" Avg_row_length="32" Data_length="32" Max_data_length="281474976710655" Index_length="2048" Data_free="0" Create_time="2009-06-18 10:02:37" Update_time="2009-06-18 10:02:43" Collation="latin1_swedish_ci" Create_options="" Comment="" />
+	</table_structure>
+	<table_data name="t1">
+	<row>
+		<field name="id">1</field>
+		<field name="text">line1
+line2
+line3</field>
+	</row>
+	</table_data>
+</database>
+</mysqldump>

mysql-test/t/loadxml.test

+#
+# Tests for "LOAD XML" - a contributed patch from Erik Wetterberg.
+#
+
+# Running the $MYSQL_DUMP tool against an embedded server does not work.
+--source include/not_embedded.inc
+
+--disable_warnings
+drop table if exists t1, t2;
+--enable_warnings
+
+create table t1 (a int, b varchar(64));
+
+
+--echo -- Load a static XML file
+load xml infile '../../std_data/loadxml.dat' into table t1
+rows identified by '<row>';
+select * from t1 order by a;
+delete from t1;
+
+
+--echo -- Load a static XML file with 'IGNORE num ROWS'
+load xml infile '../../std_data/loadxml.dat' into table t1
+rows identified by '<row>' ignore 4 rows;
+select * from t1 order by a;
+
+
+--echo -- Check 'mysqldump --xml' + 'LOAD XML' round trip
+--exec $MYSQL_DUMP --xml test t1 > "$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml" 2>&1
+delete from t1;
+--replace_result $MYSQLTEST_VARDIR MYSQLTEST_VARDIR
+--eval load xml infile '$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml' into table t1 rows identified by '<row>';
+select * from t1 order by a;
+
+--echo --Check that default row tag is '<row>
+delete from t1;
+--replace_result $MYSQLTEST_VARDIR MYSQLTEST_VARDIR
+--eval load xml infile '$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml' into table t1;
+select * from t1 order by a;
+
+--echo -- Check that 'xml' is not a keyword
+select 1 as xml;
+
+
+#
+# Bug #42520    killing load .. infile Assertion failed: ! is_set(), file .\sql_error.cc, line 8
+#
+
+--disable_query_log
+delete from t1;
+insert into t1 values (1, '12345678900987654321'), (2, 'asdfghjkl;asdfghjkl;');
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+insert into t1 select * from t1;
+--exec $MYSQL_DUMP --xml test t1 > "$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml" 2>&1
+--enable_query_log
+
+connect (addconroot, localhost, root,,);
+connection addconroot;
+create table t2(fl text);
+--let $PSEUDO_THREAD_ID=`select @@pseudo_thread_id    `
+
+--send LOAD XML LOCAL INFILE "$MYSQLTEST_VARDIR/tmp/loadxml-dump.xml"  INTO TABLE t2 ROWS IDENTIFIED BY '<person>';
+
+sleep 3;
+
+
+connection default;
+show processlist;
+
+--disable_query_log
+--eval kill $PSEUDO_THREAD_ID
+--enable_query_log
+
+disconnect addconroot;
+
+#
+# Clean up
+#
+remove_file $MYSQLTEST_VARDIR/tmp/loadxml-dump.xml;
+drop table t1;
+drop table t2;
+
+#
+# Bug #36750    LOAD XML doesn't understand new line (feed) characters in multi line text fields
+#
+
+create table t1 (
+  id int(11) not null,
+  text text,
+  primary key (id)
+) engine=MyISAM default charset=latin1;
+load xml infile '../../std_data/loadxml2.dat' into table t1;
+select * from t1;
+drop table t1;
+

sql/lex.h

@@ -607,6 +607,7 @@ static SYMBOL symbols[] = {
   { "X509",		SYM(X509_SYM)},
   { "XOR",		SYM(XOR)},
   { "XA",               SYM(XA_SYM)},
+  { "XML",              SYM(XML_SYM)}, /* LOAD XML Arnold/Erik */
   { "YEAR",		SYM(YEAR_SYM)},
   { "YEAR_MONTH",	SYM(YEAR_MONTH_SYM)},
   { "ZEROFILL",		SYM(ZEROFILL)},

sql/sql_class.cc

@@ -1602,13 +1602,17 @@ bool select_result::check_simple_select() const
 static String default_line_term("\n",default_charset_info);
 static String default_escaped("\\",default_charset_info);
 static String default_field_term("\t",default_charset_info);
+static String default_xml_row_term("<row>", default_charset_info);
 
-sql_exchange::sql_exchange(char *name,bool flag)
+sql_exchange::sql_exchange(char *name, bool flag,
+                           enum enum_filetype filetype_arg)
   :file_name(name), opt_enclosed(0), dumpfile(flag), skip_lines(0)
 {
+  filetype= filetype_arg;
   field_term= &default_field_term;
   enclosed=   line_start= &my_empty_string;
-  line_term=  &default_line_term;
+  line_term=  filetype == FILETYPE_CSV ?
+              &default_line_term : &default_xml_row_term;
   escaped=    &default_escaped;
   cs= NULL;
 }

sql/sql_class.h

@@ -89,6 +89,7 @@ enum enum_slave_exec_mode { SLAVE_EXEC_MODE_STRICT,
                             SLAVE_EXEC_MODE_LAST_BIT};
 enum enum_mark_columns
 { MARK_COLUMNS_NONE, MARK_COLUMNS_READ, MARK_COLUMNS_WRITE};
+enum enum_filetype { FILETYPE_CSV, FILETYPE_XML };
 
 extern char internal_table_name[2];
 extern char empty_c_string[1];
@@ -2364,13 +2365,15 @@ my_eof(THD *thd)
 class sql_exchange :public Sql_alloc
 {
 public:
+  enum enum_filetype filetype; /* load XML, Added by Arnold & Erik */
   char *file_name;
   String *field_term,*enclosed,*line_term,*line_start,*escaped;
   bool opt_enclosed;
   bool dumpfile;
   ulong skip_lines;
   CHARSET_INFO *cs;
-  sql_exchange(char *name,bool dumpfile_flag);
+  sql_exchange(char *name, bool dumpfile_flag,
+               enum_filetype filetype_arg= FILETYPE_CSV);
   bool escaped_given(void);
 };
 

sql/sql_yacc.yy

@@ -508,6 +508,7 @@ Item* handle_sql2003_note184_exception(THD *thd, Item* left, bool equal,
   sp_head *sphead;
   struct p_elem_val *p_elem_value;
   enum index_hint_type index_hint;
+  enum enum_filetype filetype;
   Diag_condition_item_name diag_condition_item_name;
 }
 
@@ -1116,6 +1117,7 @@ bool my_yyoverflow(short **a, YYSTYPE **b, ulong *yystacksize);
 %token  WRITE_SYM                     /* SQL-2003-N */
 %token  X509_SYM
 %token  XA_SYM
+%token  XML_SYM
 %token  XOR
 %token  YEAR_MONTH_SYM
 %token  YEAR_SYM                      /* SQL-2003-R */
@@ -1309,7 +1311,7 @@ bool my_yyoverflow(short **a, YYSTYPE **b, ulong *yystacksize);
         prepare prepare_src execute deallocate
         statement sp_suid
         sp_c_chistics sp_a_chistics sp_chistic sp_c_chistic xa
-        load_data opt_field_or_var_spec fields_or_vars opt_load_data_set_spec
+        opt_field_or_var_spec fields_or_vars opt_load_data_set_spec
         view_replace_or_algorithm view_replace
         view_algorithm view_or_trigger_or_sp_or_event
         definer_tail no_definer_tail
@@ -1338,6 +1340,7 @@ END_OF_INPUT
 %type <spname> sp_name
 %type <index_hint> index_hint_type
 %type <num> index_hint_clause
+%type <filetype> data_or_xml
 
 %type <NONE> signal_stmt resignal_stmt
 %type <diag_condition_item_name> signal_condition_information_item_name
@@ -10663,7 +10666,7 @@ use:
 /* import, export of files */
 
 load:
-          LOAD DATA_SYM
+          LOAD data_or_xml
           {
             THD *thd= YYTHD;
             LEX *lex= thd->lex;
@@ -10671,39 +10674,21 @@ load:
 
             if (lex->sphead)
             {
-              my_error(ER_SP_BADSTATEMENT, MYF(0), "LOAD DATA");
+              my_error(ER_SP_BADSTATEMENT, MYF(0), 
+                       $2 == FILETYPE_CSV ? "LOAD DATA" : "LOAD XML");
               MYSQL_YYABORT;
             }
             lex->fname_start= lip->get_ptr();
           }
-          load_data
-          {}
-        | LOAD TABLE_SYM table_ident FROM MASTER_SYM
-          {
-            LEX *lex=Lex;
-            WARN_DEPRECATED(yythd, "6.0", "LOAD TABLE FROM MASTER",
-                            "MySQL Administrator (mysqldump, mysql)");
-            if (lex->sphead)
-            {
-              my_error(ER_SP_BADSTATEMENT, MYF(0), "LOAD TABLE");
-              MYSQL_YYABORT;
-            }
-            lex->sql_command = SQLCOM_LOAD_MASTER_TABLE;
-            if (!Select->add_table_to_list(YYTHD, $3, NULL, TL_OPTION_UPDATING))
-              MYSQL_YYABORT;
-          }
-        ;
-
-load_data:
           load_data_lock opt_local INFILE TEXT_STRING_filesystem
           {
             LEX *lex=Lex;
             lex->sql_command= SQLCOM_LOAD;
-            lex->lock_option= $1;
-            lex->local_file=  $2;
+            lex->lock_option= $4;
+            lex->local_file=  $5;
             lex->duplicates= DUP_ERROR;
             lex->ignore= 0;
-            if (!(lex->exchange= new sql_exchange($4.str, 0)))
+            if (!(lex->exchange= new sql_exchange($7.str, 0, $2)))
               MYSQL_YYABORT;
           }
           opt_duplicate INTO
@@ -10713,7 +10698,7 @@ load_data:
           TABLE_SYM table_ident
           {
             LEX *lex=Lex;
-            if (!Select->add_table_to_list(YYTHD, $10, NULL, TL_OPTION_UPDATING,
+            if (!Select->add_table_to_list(YYTHD, $13, NULL, TL_OPTION_UPDATING,
                                            lex->lock_option))
               MYSQL_YYABORT;
             lex->field_list.empty();
@@ -10721,12 +10706,37 @@ load_data:
             lex->value_list.empty();
           }
           opt_load_data_charset
-          { Lex->exchange->cs= $12; }
+          { Lex->exchange->cs= $15; }
+          opt_xml_rows_identified_by
           opt_field_term opt_line_term opt_ignore_lines opt_field_or_var_spec
           opt_load_data_set_spec
           {}
-        | FROM MASTER_SYM
+        | LOAD TABLE_SYM table_ident FROM MASTER_SYM
           {
+            LEX *lex=Lex;
+            WARN_DEPRECATED(yythd, "6.0", "LOAD TABLE FROM MASTER",
+                            "MySQL Administrator (mysqldump, mysql)");
+            if (lex->sphead)
+            {
+              my_error(ER_SP_BADSTATEMENT, MYF(0), "LOAD TABLE");
+              MYSQL_YYABORT;
+            }
+            lex->sql_command = SQLCOM_LOAD_MASTER_TABLE;
+            if (!Select->add_table_to_list(YYTHD, $3, NULL, TL_OPTION_UPDATING))
+              MYSQL_YYABORT;
+          }
+        | LOAD DATA_SYM FROM MASTER_SYM
+          {
+            THD *thd= YYTHD;
+            LEX *lex= thd->lex;
+            Lex_input_stream *lip= YYLIP;
+
+            if (lex->sphead)
+            {
+              my_error(ER_SP_BADSTATEMENT, MYF(0), "LOAD DATA");
+              MYSQL_YYABORT;
+            }
+            lex->fname_start= lip->get_ptr();
             Lex->sql_command = SQLCOM_LOAD_MASTER_DATA;
             WARN_DEPRECATED(yythd, "6.0", "LOAD DATA FROM MASTER",
                             "mysqldump or future "
@@ -10734,6 +10744,11 @@ load_data:
           }
         ;
 
+data_or_xml:
+        DATA_SYM  { $$= FILETYPE_CSV; }
+        | XML_SYM { $$= FILETYPE_XML; }
+        ;
+
 opt_local:
           /* empty */ { $$=0;}
         | LOCAL_SYM { $$=1;}
@@ -10820,15 +10835,25 @@ line_term:
           }
         ;
 
+opt_xml_rows_identified_by:
+        /* empty */ { }
+        | ROWS_SYM IDENTIFIED_SYM BY text_string
+          { Lex->exchange->line_term = $4; };
+
 opt_ignore_lines:
           /* empty */
-        | IGNORE_SYM NUM LINES
+        | IGNORE_SYM NUM lines_or_rows
           {
             DBUG_ASSERT(Lex->exchange != 0);
             Lex->exchange->skip_lines= atol($2.str);
           }
         ;
 
+lines_or_rows:
+        LINES { }
+        | ROWS_SYM { }
+        ;
+
 opt_field_or_var_spec:
           /* empty */ {}
         | '(' fields_or_vars ')' {}
@@ -11924,6 +11949,7 @@ keyword_sp:
         | WEEK_SYM                 {}
         | WORK_SYM                 {}
         | X509_SYM                 {}
+        | XML_SYM                  {}
         | YEAR_SYM                 {}
         ;